{"id":"https://openalex.org/W7134074229","doi":"https://doi.org/10.48550/arxiv.2603.05487","title":"Observing and Controlling Features in Vision-Language-Action Models","display_name":"Observing and Controlling Features in Vision-Language-Action Models","publication_year":2026,"publication_date":"2026-03-05","ids":{"openalex":"https://openalex.org/W7134074229","doi":"https://doi.org/10.48550/arxiv.2603.05487"},"language":null,"primary_location":{"id":"pmh:doi:10.48550/arxiv.2603.05487","is_oa":true,"landing_page_url":null,"pdf_url":null,"source":{"id":"https://openalex.org/S4406922384","display_name":"Open MIND","issn_l":null,"issn":null,"is_oa":false,"is_in_doaj":false,"is_core":false,"host_organization":null,"host_organization_name":null,"host_organization_lineage":[],"host_organization_lineage_names":[],"type":"repository"},"license":"cc-by","license_id":"https://openalex.org/licenses/cc-by","version":"submittedVersion","is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"Article"},"type":"preprint","indexed_in":["datacite"],"open_access":{"is_oa":true,"oa_status":"green","oa_url":null,"any_repository_has_fulltext":true},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5128240530","display_name":"Hugo Buurmeijer","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Buurmeijer, Hugo","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":null,"display_name":"Alonso, Carmen Amo","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Alonso, Carmen Amo","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5114159472","display_name":"Aiden Swann","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Swann, Aiden","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"last","author":{"id":"https://openalex.org/A5128222191","display_name":"Marco Pavone","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Pavone, Marco","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]}],"institutions":[],"countries_distinct_count":0,"institutions_distinct_count":0,"corresponding_author_ids":[],"corresponding_institution_ids":[],"apc_list":null,"apc_paid":null,"fwci":null,"has_fulltext":false,"cited_by_count":0,"citation_normalized_percentile":null,"cited_by_percentile_year":null,"biblio":{"volume":null,"issue":null,"first_page":null,"last_page":null},"is_retracted":false,"is_paratext":false,"is_xpac":false,"primary_topic":{"id":"https://openalex.org/T11714","display_name":"Multimodal Machine Learning Applications","score":0.8353000283241272,"subfield":{"id":"https://openalex.org/subfields/1707","display_name":"Computer Vision and Pattern Recognition"},"field":{"id":"https://openalex.org/fields/17","display_name":"Computer Science"},"domain":{"id":"https://openalex.org/domains/3","display_name":"Physical Sciences"}},"topics":[{"id":"https://openalex.org/T11714","display_name":"Multimodal Machine Learning Applications","score":0.8353000283241272,"subfield":{"id":"https://openalex.org/subfields/1707","display_name":"Computer Vision and Pattern Recognition"},"field":{"id":"https://openalex.org/fields/17","display_name":"Computer Science"},"domain":{"id":"https://openalex.org/domains/3","display_name":"Physical Sciences"}},{"id":"https://openalex.org/T12090","display_name":"Language and cultural evolution","score":0.0284000001847744,"subfield":{"id":"https://openalex.org/subfields/3316","display_name":"Cultural Studies"},"field":{"id":"https://openalex.org/fields/33","display_name":"Social Sciences"},"domain":{"id":"https://openalex.org/domains/2","display_name":"Social Sciences"}},{"id":"https://openalex.org/T11307","display_name":"Domain Adaptation and Few-Shot Learning","score":0.017799999564886093,"subfield":{"id":"https://openalex.org/subfields/1702","display_name":"Artificial Intelligence"},"field":{"id":"https://openalex.org/fields/17","display_name":"Computer Science"},"domain":{"id":"https://openalex.org/domains/3","display_name":"Physical Sciences"}}],"keywords":[{"id":"https://openalex.org/keywords/interpretability","display_name":"Interpretability","score":0.907800018787384},{"id":"https://openalex.org/keywords/representation","display_name":"Representation (politics)","score":0.5848000049591064},{"id":"https://openalex.org/keywords/domain-adaptation","display_name":"Domain adaptation","score":0.5590000152587891},{"id":"https://openalex.org/keywords/task","display_name":"Task (project management)","score":0.4537999927997589},{"id":"https://openalex.org/keywords/internal-model","display_name":"Internal model","score":0.4075999855995178},{"id":"https://openalex.org/keywords/transformer","display_name":"Transformer","score":0.400299996137619},{"id":"https://openalex.org/keywords/interval","display_name":"Interval (graph theory)","score":0.35899999737739563}],"concepts":[{"id":"https://openalex.org/C2781067378","wikidata":"https://www.wikidata.org/wiki/Q17027399","display_name":"Interpretability","level":2,"score":0.907800018787384},{"id":"https://openalex.org/C2776359362","wikidata":"https://www.wikidata.org/wiki/Q2145286","display_name":"Representation (politics)","level":3,"score":0.5848000049591064},{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.5837000012397766},{"id":"https://openalex.org/C2776434776","wikidata":"https://www.wikidata.org/wiki/Q19246213","display_name":"Domain adaptation","level":3,"score":0.5590000152587891},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.46480000019073486},{"id":"https://openalex.org/C2780451532","wikidata":"https://www.wikidata.org/wiki/Q759676","display_name":"Task (project management)","level":2,"score":0.4537999927997589},{"id":"https://openalex.org/C28427503","wikidata":"https://www.wikidata.org/wiki/Q13580300","display_name":"Internal model","level":3,"score":0.4075999855995178},{"id":"https://openalex.org/C66322947","wikidata":"https://www.wikidata.org/wiki/Q11658","display_name":"Transformer","level":3,"score":0.400299996137619},{"id":"https://openalex.org/C2778067643","wikidata":"https://www.wikidata.org/wiki/Q166507","display_name":"Interval (graph theory)","level":2,"score":0.35899999737739563},{"id":"https://openalex.org/C163175372","wikidata":"https://www.wikidata.org/wiki/Q3339222","display_name":"Linear model","level":2,"score":0.321399986743927},{"id":"https://openalex.org/C119857082","wikidata":"https://www.wikidata.org/wiki/Q2539","display_name":"Machine learning","level":1,"score":0.3010999858379364},{"id":"https://openalex.org/C123657996","wikidata":"https://www.wikidata.org/wiki/Q12271","display_name":"Architecture","level":2,"score":0.29429998993873596},{"id":"https://openalex.org/C139807058","wikidata":"https://www.wikidata.org/wiki/Q352374","display_name":"Adaptation (eye)","level":2,"score":0.290800005197525},{"id":"https://openalex.org/C48209547","wikidata":"https://www.wikidata.org/wiki/Q1331104","display_name":"Controllability","level":2,"score":0.2906999886035919},{"id":"https://openalex.org/C36503486","wikidata":"https://www.wikidata.org/wiki/Q11235244","display_name":"Domain (mathematical analysis)","level":2,"score":0.2793000042438507},{"id":"https://openalex.org/C100609095","wikidata":"https://www.wikidata.org/wiki/Q1335050","display_name":"Embodied cognition","level":2,"score":0.26499998569488525},{"id":"https://openalex.org/C80444323","wikidata":"https://www.wikidata.org/wiki/Q2878974","display_name":"Theoretical computer science","level":1,"score":0.2630000114440918},{"id":"https://openalex.org/C2776401178","wikidata":"https://www.wikidata.org/wiki/Q12050496","display_name":"Feature (linguistics)","level":2,"score":0.2565000057220459},{"id":"https://openalex.org/C204323151","wikidata":"https://www.wikidata.org/wiki/Q905424","display_name":"Range (aeronautics)","level":2,"score":0.25119999051094055}],"mesh":[],"locations_count":2,"locations":[{"id":"pmh:doi:10.48550/arxiv.2603.05487","is_oa":true,"landing_page_url":null,"pdf_url":null,"source":{"id":"https://openalex.org/S4406922384","display_name":"Open MIND","issn_l":null,"issn":null,"is_oa":false,"is_in_doaj":false,"is_core":false,"host_organization":null,"host_organization_name":null,"host_organization_lineage":[],"host_organization_lineage_names":[],"type":"repository"},"license":"cc-by","license_id":"https://openalex.org/licenses/cc-by","version":"submittedVersion","is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"Article"},{"id":"doi:10.48550/arxiv.2603.05487","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2603.05487","pdf_url":null,"source":{"id":"https://openalex.org/S4306400194","display_name":"arXiv (Cornell University)","issn_l":null,"issn":null,"is_oa":true,"is_in_doaj":false,"is_core":false,"host_organization":"https://openalex.org/I205783295","host_organization_name":"Cornell University","host_organization_lineage":["https://openalex.org/I205783295"],"host_organization_lineage_names":[],"type":"repository"},"license":"cc-by","license_id":"https://openalex.org/licenses/cc-by","version":null,"is_accepted":false,"is_published":null,"raw_source_name":null,"raw_type":"Preprint"}],"best_oa_location":{"id":"pmh:doi:10.48550/arxiv.2603.05487","is_oa":true,"landing_page_url":null,"pdf_url":null,"source":{"id":"https://openalex.org/S4406922384","display_name":"Open MIND","issn_l":null,"issn":null,"is_oa":false,"is_in_doaj":false,"is_core":false,"host_organization":null,"host_organization_name":null,"host_organization_lineage":[],"host_organization_lineage_names":[],"type":"repository"},"license":"cc-by","license_id":"https://openalex.org/licenses/cc-by","version":"submittedVersion","is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"Article"},"sustainable_development_goals":[],"awards":[],"funders":[],"has_content":{"grobid_xml":false,"pdf":false},"content_urls":null,"referenced_works_count":0,"referenced_works":[],"related_works":[],"abstract_inverted_index":{"Vision-Language-Action":[0],"Models":[1,19],"(VLAs)":[2],"have":[3],"shown":[4],"remarkable":[5],"progress":[6],"towards":[7,137],"embodied":[8],"intelligence.":[9],"While":[10],"their":[11,27,61],"architecture":[12],"partially":[13],"resembles":[14],"that":[15,96,144,170],"of":[16,34,42,112],"Large":[17],"Language":[18],"(LLMs),":[20],"VLAs":[21,171],"exhibit":[22],"higher":[23],"complexity":[24],"due":[25],"to":[26,60,68,76,127,177],"multi-modal":[28],"inputs/outputs":[29],"and":[30,36,82,88,103,132,165,188],"often":[31],"hybrid":[32],"nature":[33],"transformer":[35],"diffusion":[37],"heads.":[38],"This":[39],"is":[40],"part":[41],"the":[43,55,134],"reason":[44],"why":[45],"insights":[46],"from":[47],"mechanistic":[48],"interpretability":[49],"in":[50,100,124],"LLMs,":[51],"which":[52],"explain":[53],"how":[54,105],"internal":[56,130,174],"model":[57],"representations":[58,131],"relate":[59],"output":[62,136],"behavior,":[63],"do":[64],"not":[65],"trivially":[66],"transfer":[67],"VLA":[69,162],"counterparts.":[70],"In":[71,90],"this":[72,78],"work,":[73],"we":[74,92,117],"propose":[75],"close":[77],"gap":[79],"by":[80,110],"introducing":[81],"analyzing":[83],"two":[84],"main":[85],"concepts:":[86],"feature-observability":[87],"feature-controllability.":[89],"particular,":[91],"first":[93],"study":[94],"features":[95],"are":[97],"linearly":[98],"encoded":[99],"representation":[101],"space,":[102],"show":[104,143],"they":[106],"can":[107,148],"be":[108],"observed":[109],"means":[111],"a":[113,119,138,151],"linear":[114,121],"classifier.":[115],"Then,":[116],"use":[118],"minimal":[120],"intervention":[122],"grounded":[123],"optimal":[125],"control":[126],"accurately":[128],"place":[129],"steer":[133,150],"VLA's":[135],"desired":[139],"region.":[140],"Our":[141],"results":[142],"targeted,":[145],"lightweight":[146],"interventions":[147],"reliably":[149],"robot's":[152],"behavior":[153],"while":[154],"preserving":[155],"closed-loop":[156],"capabilities.":[157],"We":[158],"demonstrate":[159],"on":[160],"different":[161],"architectures":[163],"($\u03c0_{0.5}$":[164],"OpenVLA)":[166],"through":[167],"simulation":[168],"experiments":[169],"possess":[172],"interpretable":[173],"structure":[175],"amenable":[176],"online":[178],"adaptation":[179],"without":[180],"fine-tuning,":[181],"enabling":[182],"real-time":[183],"alignment":[184],"with":[185],"user":[186],"preferences":[187],"task":[189],"requirements.":[190]},"counts_by_year":[],"updated_date":"2026-07-01T06:00:48.157686","created_date":"2026-03-07T00:00:00"}
